yahoo-eng-team team mailing list archive

[Pierre Libeau <1939116@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

Description
===========
On the live snapshot feature to have data consistency, the instance can be boot with image property hw_qemu_guest_agent set to yes and QEMU agent install in the image. The instance with local storage creates a disk mirror to have a copy and the disk. With property hw_qemu_guest_agent the compute contacts the instance to freeze the filesystem before starting the mirror and contacts the instance again after the end of mirroring to unfreeze the filesystem. The issue is for instance with a "big" disk the filesystem is frozen for a long time and the impact is instance hung "task blocked for more than 120 seconds."

Steps to reproduce
==================
1/ Create an image with property hw_qemu_guest_agent=yes and QEMU agent install in the image
2/ Create an instance with local storage (disk size 200Go) with real data on the disk
3/ Start live-snapshot
4/ Instance frozen during the mirroring


Expected result
===============
Nova compute can wait until the end of mirroring to freeze the file system in the instance.

Actual result
=============
The instance is frozen during the mirroring

** Affects: nova
     Importance: Undecided
     Assignee: Pierre Libeau (pierre-libeau)
         Status: New

** Changed in: nova
     Assignee: (unassigned) => Pierre Libeau (pierre-libeau)

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1939116

Title:
  File system freeze before end of mirroring generate kernel task
  blocked for more than 120 seconds

Status in OpenStack Compute (nova):
  New

Bug description:
  Description
  ===========
  On the live snapshot feature to have data consistency, the instance can be boot with image property hw_qemu_guest_agent set to yes and QEMU agent install in the image. The instance with local storage creates a disk mirror to have a copy and the disk. With property hw_qemu_guest_agent the compute contacts the instance to freeze the filesystem before starting the mirror and contacts the instance again after the end of mirroring to unfreeze the filesystem. The issue is for instance with a "big" disk the filesystem is frozen for a long time and the impact is instance hung "task blocked for more than 120 seconds."

  Steps to reproduce
  ==================
  1/ Create an image with property hw_qemu_guest_agent=yes and QEMU agent install in the image
  2/ Create an instance with local storage (disk size 200Go) with real data on the disk
  3/ Start live-snapshot
  4/ Instance frozen during the mirroring

  
  Expected result
  ===============
  Nova compute can wait until the end of mirroring to freeze the file system in the instance.

  Actual result
  =============
  The instance is frozen during the mirroring

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1939116/+subscriptions